read-logs

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs the model to read raw run logs (which the prompt warns routinely contain auth/refresh/push tokens, cookies, emails, etc.) into context and to surface/log lines as needed, so the model will handle secret values and may be required to reproduce them verbatim.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This skill intentionally captures and forwards runtime logs (which commonly contain auth/refresh tokens, emails, and customer data) to the model and also offers to modify editor run settings to persistently capture those logs — an explicit data‑exfiltration / credential‑leak risk that could be abused.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). Required workflow reads the app’s runtime run log from /tmp/flutter-<repo>.log (captured from flutter run output via script transcript or DAP dart.dapLogFile), which is outsider-authored free text in the sense that it includes arbitrary console output/messages produced by the running app and its environment, and that text is ingested into the agent’s LLM context when the skill loads $CLEAN.